Before I read your paper, I didn't realize the power of hypothesis tests. I was surprised to see theoretical bounds on anomaly detection with such weak constraint through simple Chi-Squared goodness-of-fit. Delicate idea!